Client - Media - PureDisk Linux network traffic
Hello everyone. We have a backup solution and I am trying to find a bottleneck. Configuration is as follows: Veritas SFHA clustered Master server running NBU 7.5.0.6 on RHEL6.1. Number of Media servers. And 3 pairs of PureDisk 6.5.5pools as a storage servers. My concern is: How does network traffic flow during backup assuming clientside deduplication is enabled? I tried tcpdump and appears that it is client - media - puredisk and vica-versa. But there arenbstproxy processes. Appreciate on any assistance. Thank you.Solved4.1KViews0likes2Comments
About Pure Disk
Hi All, Kindly let me know the following few questions about pure disk. 1)what is pure disk? 2)About pure disk architecture? 3)About pure disk deduplication? 4)About pure disk web user interface? 5)Can i know the role of vssadminlist writers related to system state backup failures in pure disk.Solved3.8KViews3likes2Comments
CLI Command For Dedup Rates
NBU Experts, Do any of you know if there is a CLI command that can tell the dedup rate for all the images created, and stored in a PureDisk pool? For examaple, if I run bpimagelist on the master server I get some good info, but not the dedup rate for each image. Thanks, AVazquez2.6KViews0likes7CommentsMSDP question re full at 96% vs. high water mark at 98%
When does full disk processing take place? At 96% or 98% ? And doesn't this all imply that a high water mark is probably best set to something like 90% (max)or probably (85%) so that clenup can occur before it gets full. If so, then why is teh default 98% Page 68 of the NetBackup v7.5 Deduplication Guide: http://www.symantec.com/business/support/index?page=content&id=DOC5187 ...says this: About deduplication storage capacity The maximum deduplication storage capacity is 64 TBs. NetBackup reserves 4 percent of the storage space for the deduplication database and transaction logs. Therefore, a storage full condition is triggered at a 96 percent threshold. For performance optimization, Symantec recommends that you use a separate disk, volume, partition, or spindle for the deduplication database. If you use separate storage for the deduplication database, NetBackup still uses the 96 percent threshold to protect the data storage from any possible overload. If your storage requirements exceed the capacity of a media server deduplication node, do one of the following: - Use more than one media server deduplication node. - Use a PureDisk storage pool as the storage destination. A PureDisk storage pool provides larger storage capacity; PureDisk also provides global deduplication. See “About the deduplication storage destination” on page 22. Only one deduplication storage path can exist on a media server. You cannot add another storage path to increase capacity beyond 64 TBs. ...but says this on page 82: High water mark The High water mark setting is a threshold that invokes the following actions: - The High water mark indicates that the PureDiskVolume is full. When the PureDiskVolume reaches the High water mark, NetBackup fails any backup jobs that are assigned to the storage unit. NetBackup also does not assign new jobs to a storage unit in which the disk pool is full. NetBackup also fails backup jobs if the PureDiskVolume does not contain enough storage for its estimated space requirement. - NetBackup begins image cleanup when the PureDiskVolume reaches the High water mark; image cleanup expires the images that are no longer valid. NetBackup again assigns jobs to the storage unit when image cleanup reduces the PureDiskVolume capacity to less than the High water mark. The default is 98%.2.1KViews0likes1CommentHow to determine if CR rebasing completed
Hi, I am trying to find a way to find out if the post 6.6.3a PureDisk upgrade Content Router rebasing has completed. It is quite easy to check with the crcontrol if the rebasing option is enabled or disabled at all but I would like to know if the rebasing is still "in progress" or is this "completed". I am aware that this rebasing runs in backgroud and not 24/7 but how to check if all CR entries have been rebased?? Thanks!1.8KViews1like4CommentsMSDP encryption (ServerOptions=agent_crypt) (or should it be =encrypt)
Hi Forum, Page 30 of the NetBackup Deduplication Guide v7.5 says that... by default MSDP encryption is disabled by default, but they it is recommended by Symantec. Page 70 says that to enable encryption that we need to add "agent_crypt" to "ServerOptions=" in the ContentRouter.cfg file. But looking at the comments in the file: ; [no]agent_crypt : always disable/enable client-side encryption; ; [no]encrypt : always disable/enable encryption; ...indicates that using "agent_crypt" instructs the client-side deduplication agent to encrypt data. At our site we are not implementing client side deduplication so it woudl seem to me that setting :agent_crypt" won't have any effect. My questions are: 1) Is "agent_crypt" really the option to use if what we are trying to achieve is encryption of the post-dedupe blocks on teh MSDP data storage local to teh MSDP media server? 2) What does the "ServerOptions=encrypt" option do, and how is it different to "agent_crypt" ? 3) If we can determine which option really does enable post-dedupe encryption - then does anyone know what the CPU overhead is ? 4) Are there any particular models of CPU (SPARC or Intel or AMD) that implement the encryption/decryption in their hardware instruction sets - i.e. are there any CPUs that are much better suited for use with MSDP encryption ? Thanks, Dave.Solved1.8KViews1like4CommentsPuredisk Replication Import performance
Having an issue with Puredisk taking forever to import replicated data selections. The Import jobs get queued up into the 1000's. The replication jobs themeselves are completing fine- the data import is what's getting delayed. Just curious what the primary performance constraints are with this process- should I be looking at CPU, RAM, Disk? I have an all-in-one system as the replication target- should I look at spinning off one of the services to it's own box? And if so, which one should I be looking at? Not looking for a definitive "your problem is X". more like "Data Imports are (RAM/CPU/etc) intensive, so make sure you have plenty of that"Solved1.6KViews0likes3Comments
SLP Duplication Behaviour
Master Server ( nbu01 ) Windows 2008 NetBackup Master 7.5.0.7 Media Server ( nbu02 ) Window 2008 NetBackup Media 7.5.0.7 MSDP with puredisk Setup storage server load balancing for dedupe pool nbu01 and nbu02 setup SLP backup to disk then duplicate to tape both nbu01 and nbu02 connected to the same tape library ofcourse with SSO enable All backup and duplication went well but found out that during duplication to tape, always server nbu01 trigger to job even the tape storage unit is connected to nbu02. I want to know if duplication job trigger from nbu01, let say backup image is 100GB, from the dedupe guide, it say that it will always rehydrate the data where the dedupe pool server is, so means nbu02, so do this means that it will send 100GB data to nbu01, then nbu01 will send back 100GB of data back to nbu02, because i select tape storage unit that connected to the nbu02, and write to tape. If above is true, means my bottleneck will be at the LAN, and how do i force only use nbu02 during duplication job. I also planning to remove the load balancing, since i dont see much help on the backup and duplication, but i still have job queuing for duplication. If i inactive those job, remove the nbu01 from the load balancing, and activate backup duplication, will the duplication job still look for nbu01 ? or i need to cancel all job, and do manual duplication ? anyone please help.1.3KViews0likes2CommentsAny recommendations for volume manager and file system type for MSDP on Linux...
Hi Forum, I've had a quick search through the DeDupe Admin Guide for v7.5 and I couldn't find any recommendations for volume manager or file system type for the MSDP database and/or data spaces. I guess an appliances uses VxVM and VxFS - but what else might be suitable if we were to build our own Linux based MSDP media server. The local server disks will re RAID-1, and the SAN LUNs will be: - MSDP DB on one x 2 TB LUN from 3PAR - MDDP data on four x 16 TB LUNs from 3PAR (i.e. 4 x PG of RAID-6 6D-2P) I guess my question really relates to the MSDP data space - I was thinking of a simple logical volume of the four LUNs concat together and an ext4 file system on that logical volume. Or might I be better off in carving the MSDP data storage as a concat 2 x LUNs (i.e. 2 x PG of RAID-6 14D-2P)? And I assume a sector cluster/blocking factor of 64 KB is appropriate for ext4, just as it is for NTFS. Anyone have any advice, or willing to share what they did and whether they ripped anything up and started again?1.2KViews0likes2Comments